YOLOv7-P: a lighter and more effective UAV aerial photography object detection algorithm

Fengxi Sun,Ning He,Xin Wang,Hongfei Liu,Yuxiang Zou
DOI: https://doi.org/10.1007/s11760-024-03476-8
IF: 1.583
2024-09-15
Signal Image and Video Processing
Abstract:Because of the special way an unmanned aerial vehicle (UAV) acquires aerial photography, UAV images have the characteristics of large coverage area, complex background, and a large proportion of small targets, which exacerbate the difficulty of object detection. Additionally, UAV-based aerial image detection needs to meet lightweight and real-time capabilities. To address these issues, this paper proposes a lightweight model YOLOv7-P that is based on YOLOv7 but has a stronger detection capability for small targets. First, partial convolution (PConv) is used to reduce redundant parameters and computation in YOLOv7. Second, an optimal combination of detection heads is determined that can significantly improve the detection performance of small objects. Third, a novel lightweight convolution called PConv-wide is proposed to replace RepConv in the network, thus simplifying the network without affecting detection accuracy. Finally, the normalized wasserstein distance loss is reasonably combined with the complete intersection over union loss to further improve the sensitivity of the network to small targets. The proposed YOLOv7-P model strikes a delicate balance between precision and parameter count. Compared with the baseline YOLOv7 network, it reduces parameter count by 47.1% without increasing computational complexity and boosts AP50 by 8% and mAP by 5.4% on the VisDrone dataset.
engineering, electrical & electronic,imaging science & photographic technology
What problem does this paper attempt to address?